bespoke data solutions for your AI models

We are the world's first data cooperative enabling economic opportunities for rural Indians.

Over 3 million paid digital tasks have been completed on the Karya platform

Our goal at Karya Inc. is simple. We want to bring dignified, digital work to rural Indians. By providing high-quality data annotation services for AI/ML models to our clients, we can create economic opportunities to people in rural communities.

Here are a few things our workers can do for you immediately:

Building Speech Datasets

Our platform is currently being used by Gates Foundation to build language models in 10 Indian states. We are also working with Microsoft to build large-scale text corpuses, and record speech data in major Indian languages. Karya workers have also built conversational speech datasets in English and other languages.

Document Digitisation

Karya workers can digitise documents in English and in local Indian languages. We wrote a research paper describing our work in document digitisation which we presented at the reputed CHI conference in Glasgow. You can read our research papers in the "Research" section below.

Image Datasets

Karya workers can help you collect images to build a diverse image corpus. We recently worked with Microsoft to collect 10,000 images of Bengali signboards. This allowed Microsoft to build an OCR technology that could identify the text in signboards and translate them to English.

Image Annotation

Karya workers can help you annotate and label images, create bounding boxes around relevant items within an image, and help you classify images. Our workers are currently working on annotating a corpus of 100,000 newspaper photos.

Our state of the artapplication andplatform makes everything easy for you

Simply upload the task on our platform, identify your key metrics, and our platform will distribute your tasks to our workers across India.

How do we start?

How it works

1

Reach out to us with an ask

Simply drop us a note using the Contact page/ email us at thekaryainc@gmail.com. One of our team members will get back to you with ways we can fulfill your query.

2

We get the work done, promising high accuracy

We start by selecting a list of villages that meet your criteria. With the help of our on-ground team, we deploy your work. We ensure the highest standards of acccuracy, and take care of everything.

3

You get your work, and workers in rural India move out of poverty

We provide you your results, in time, and with our standard high accuracy rates. Our participants in rural India get paid, and they are able to enjoy a dignified livelihood. A genuine win-win!

dive into our research papers

We have designed a crowdsourcing platform that is inclusive and accessible to users in rural communities. Our platform provides work and instructions in the regional language of the users. Our two-tier server architecture allows anyone with a smartphone including those with no data connectivity to participate on our platform, receive work, and get paid. Read our publications below:

Exploring Crowdsourced Work in Low-Resource Settings: (CHI) 2019
Crowdsourcing Speech Data for Low-Resource Languages from Low-Income Workers: (LREC) 2020
109 hours of Marathi Speech Data (Open-Sourced and Collected by Karya workers)

Start giving work with Karya

Simply drop your email below, and one of our team members will reach out to you. It's that easy.

Thank you! Your submission has been received!
Oops! Something went wrong while submitting the form.